Instructor Info: Dave Tucker, LSSMBB ProModel Senior Consultant Office:

Instructor Info: Dave Tucker, LSSMBB ProModel Senior Consultant Office: 321.567.5642 dtucker@promodel.com 1

Understand how simulation models can be utilized within the DMAIC approach Examine a completed model for the application of Lean Concepts to an as is process Identify & eliminate waste in a process Perform analysis of model data Build Scenarios using Lean Concepts 2 2

Tools Tasks Gates (also known as Tollgates) Define customers and requirements Define Sponsor Define Process Owner and other Stakeholders Develop the Problem Statement, Goals, Benefits Define Measures Determine Root Causes Define Potential Solutions Define Control Plan Validate the Measurement System Identify Value Add versus Non- Value Add Process Steps Assess Potential Solutions Define the Data Collection Plan Identify sources of Variation Develop Proposed Solution Collect the Data Define Team, Define X s and the Big Y Develop Project Plan Develop High Level Process Map Determine if Process is in Control Define the Process Capability Develop the relationship between Y and critical few X s Determine the improvement opportunities to apply Lean Pilot the Solution Define to be Process Map Validate the Potential Improvement Project Charter Value Stream Map Run Chart/Time Series Brainstorming/Affinity Diagram Control Plan Develop Training Materials, Processes, Procedures Implement Statistical Process Controls Implement Visual Controls and Poka-Yoke Determine Process Capability Verify Benefits, Cost Savings, Finalize Documentation Transition to Process Owner, Close Project, and Celebrate ROI/Payback Analysis * Simulation Ishikawa Diagram/5 Whys * Simulation Process Capability Voice of the Customer Measurement System Analysis Cause and Effect Matrix FMEA Process Sigma Calculator Voice of the Business Data Collection Plan Regression Analysis Design of Experiments Mistake Proofing Stakeholder Analysis Check Sheets Non-Parametric Analysis Piloting Visual Process Control Communication Plan Benchmarking * Simulation Pugh Matrix * Simulation RACI Value of Speed (Little s Law) ANOVA Process Balancing SIPOC Statistical Sampling Components of Variation Analytical Batch Sizing * Simulation Pareto Charts Hypothesis Testing Process Flow Improvement Kano Analysis Control Charts NVA Analysis Kanban Process Capability Queuing Theory Stocking Strategy Histograms FMEA Force Field Analysis 3 3

Paint Shop Work Unit 2 Work Unit Primer Oven_1 Load_Fixture O1 Buffer Fixture Clean & Prep Fixture Rework Paint 1 Paint 2 Worker 1 Redo Worker 2 Inspection Unload Fixture Oven 2 O2 Buffer Worker 3 (Exit) 0000 0000 0000 0000 Number Completed Number Rejected Work in Process Cycle Time (minutes) 5 5

Problem Statement Current average Cycle Time (CT) is ~642 minutes; customers are complaining that they need the product sooner; employees are complaining that some are working too hard Voice of the Customer / Voice of the Business Customer survey says they want CT to be <= 300 minutes Business is not sure if that CT can be met; however, they want orders quickly filled As is Process Flow 6 6

1 Paint Shop Layout 5 5 Incoming Roller Table Primer Table Table Paint 1 Table Paint 2 Table 4 Door Load Fixture Supplies Outgoing 1 0 Bench Worker 3 Rework Area Inspection Area Bench Worker 1 Worker 2 9 8 Fixture Clean & Prep Table 3 Door 7 Unload Fixture Table 6 Oven 2 Bank of 4 Ovens Table Oven 1 2 7 7

Define Measures Collect the Data Define X s and the Big Y Determine if Process is in Control Define the Process Capability 9 9

TP is ~331 WIP is Steadily Increasing 10 10

CT is Increasing 11 11

CT Averages ~642 Minutes Redo CT s are Double 12 12

Entities are Blocked ~68% of the time Entities in Operation ~19% of the time 13 13

Workers are Utilized ~62, 78 & 37% of the time 14 14

Blockage at Paint Buffers & Load Fixture 15 15

Most Buffers Have Entities Throughout Run 16 16

1. Select the Six Sigma Analysis Feature 2. Add the Parameter for the Variable, Cycle Time 3. Pick Observation for Stat Collection 4. Enter an Upper Spec. Limit of 300 5. Hit the Launch Minitab button 17 17

vcycletime (Baseline) - Capability Analysis (using 95.0% confidence) CT is Not Capable P rocess Data LSL 0 Target * USL 300 Sample M ean 605.789 Sample N 7500 StDev (Within) 122.496 StDev (O v erall) 249.586 LSL USL Within Overall P otential (Within) C apability Z.Bench -2.50 Low er C L -2.55 Z.LSL 4.95 Z.U SL -2.50 C pk -0.83 Low er C L -0.85 Upper C L -0.82 O v erall C apability O bserv ed Performance PPM < LSL 0.00 PPM > USL 882133.33 PPM Total 882133.33 0 275 550 Exp. Within P erformance PPM < LSL 0.38 PPM > USL 993725.43 PPM Total 993725.81 825 1100 1375 Exp. O v erall Performance PPM < LSL 7608.52 PPM > USL 889746.92 PPM Total 897355.44 1650 1925 Z.Bench -1.27 Low er C L -1.31 Z.LSL 2.43 Z.U SL -1.23 Ppk -0.41 Low er C L -0.42 Upper C L -0.40 C pm * Low er C L * 18 18

A review of the Operation Times reveals that the Paint Booths have the longest time per part. Furthermore, the Booths have only a single part capacity which causes backups. The Batching and long operation times at both Ovens will also create delays. Time Avg Time Batch Time per In Out Step Time Unit (min) Size B4 Part (min) Capacity Buffer Buffer Resource Primer T(3, 4, 5) min 4 1 4 1 999 0 -- Oven_1 60 min 60 15 4 1 - B 999 999 -- Load_Fixture N(60, 10) sec 1 1 1 1 0 0 Worker 2 Fixture_Clean_Prep U(2, 1) min 2 1 2 1 999 0 -- Paint_1 N(8, 1) min 8 1 8 1 1 0 -- Paint_2 N(8, 1) min 8 1 8 1 1 0 -- Oven_2 40 min 40 10 4 4 - B 100 100 -- Unload_Fixture N(50, 10) sec 0.83 1 0.83 1 999 0 Worker 2 Inspection N(3, 0.1) min 3 1 3 1 999 0 Worker 1 Rework T(2, 7, 15) min 7 1 7 1 999 0 Worker 2 19 19

WIP and CT are increasing which indicates one or more process constraints Some Entity blockage occurs at several Activities Resource utilization is out of balance Process Cycle Time is statistically out-of-control and therefore, inconsistent Process is not capable of achieving customer specifications of <= 300 CT minutes Process flow is unbalanced and not level-loaded 20 20

Determine the improvement opportunities to apply Lean concepts Identify Value Add versus Non-Value Add Process Steps Develop the relationship between Y and critical few X s Determine Root Causes 22 22

Eliminating Waste Bottleneck Identification Queue Reduction Equipment Setup Reduction Building Pull Systems Process Flow Improvement 23 23

1. Unnecessary material handling or Transportation 2. Excess Inventory (just in case) 3. Excess or inefficient operator Motion 4. Waiting for materials or resources 5. Overproduction (often causes the other types of wastes) 6. Over processing / Unnecessary steps 7. Production of Defects (any type) 24 24

Using Customer Orders to Trigger Production Will Eliminate Waste of Over-Production (5) Reduced Operation Time Could Eliminate Waste of Waiting (4) Paint Shop Different Batch Sizes Could Eliminate Waste of Waiting (4) Cellular Layout Could Eliminate Waste of Excess Operator Motion to / fm Load & Unload (3) Work Unit 2 Work Unit Primer Oven_1 Load_Fixture O1 Buffer Fixture Clean & Prep Fixture Rework Paint 1 Paint 2 Reduced Operation Time Could Eliminate Waste of Waiting (4) Worker 3 Worker 1 Using Worker Resources Differently Could Eliminate Waste of Waiting (4) Inspection Redo Unload Fixture Worker 2 O2 Buffer Oven 2 Different Batch Sizes Could Eliminate Waste of Waiting (4) Cellular Layout Could Eliminate Waste of Transportation (1) Improved Quality Could Eliminate Waste of Inspection & Defects (7) = NVA Step (Exit) 0000 Number Completed 0000 Number Rejected 0000 Work in Process 0000 Cycle Time (minutes) 25 25

Eliminating Waste Bottleneck Identification Queue Reduction Equipment Setup Reduction Building Pull Systems Process Flow Improvement 26 26

Q Q Paint Shop Q Work Unit Work Unit 2 Primer Oven_1 Load_Fixture O1 Buffer Fixture Clean & Prep Q Fixture Rework Paint 1 Paint 2 Legend: Worker 1 Worker 3 Q Inspection (Exit) Redo Q Worker 2 Unload Fixture 0000 Number Completed Q O2 Buffer 0000 Number Rejected Oven 2 0000 Q Work in Process 0000 Cycle Time (minutes) Q = Setup Reduction = Bottleneck = Queue Reduction = Pull System 27 27

Waste Over-production Waiting Operator Motion Transportation Defects & Inspection Bottlenecks At Load Fixture At Paint 1 & 2 No Pull Systems To Primer To Load Fixture Queues Primer Oven 1 & 2 Load & Unload Fixture Fixture Clean & Prep Inspection & Rework Setups Paint 1 & 2 Traditional Shop Layout Functional not Cellular Other 28 28

1 Primer 3 Fixture Clean 5 Paint 1 & 2 7 Unload Fixture 2 Queue Waste Over-production Waste Waiting Limited Capacity Queue Batch Size Oven 1 No Pull System Waste Op Time Waiting Limited Capacity Bottleneck Queue 4 Queue Waste Waiting Limited Resources Waste Operator Motion Limited Capacity Waste Transportation Use of Resources Limited Capacity Load Fixture Waste Waiting No Pull 6 Queue Batch Size Setups Bottleneck Limited Capacity Oven 2 Limited Capacity Waste Op Time Waiting Waste Defects 8 9 Queue Queue Limited Resources Waste Waiting Use of Resources Quality Problems Waste Waiting Insp / Rework Limited Resources Use of Resources Waste Operator Motion Waste Operator Motion Waste Waiting Use of Resources Functional Layout General - Shop Cycle Time Too High Continuous Part Order Arrivals Equipment with Limited Capacities Unbalanced Process Flow Rework Rates 29 29

What is the effect of Batching, Capacity, Queuing and Operation times at Ovens 1 & 2? We are told that the total time to get a part through those Ovens is about the same even though Batching, Capacity and Operation times are different. Is the waiting and operation total time about the same at both of the ovens? Paint Shop Work Unit Batch Size: 15 Op Time: 60 min Capacity: 1 B Work Unit 2 Worker 1 Worker 3 Q Q Primer Oven_1 Load_Fixture Rework Inspection (Exit) Q Redo Worker 2 Q Unload Fixture 0000 Number Q O1 Buffer Q O2 Buffer Paint 1 0000 Number Oven 2 Paint 2 0000 Q Work in Fixture Clean & Prep 0000 Cycle Time Q Fixture Q Legend: = Setup Reduction Batch Size: 10 Op Time: 40 min = Bottleneck Capacity: 4 B = Queue Reduction = Pull System 30 30

1. Create 2 Attributes & 2 Variables 2. Create logic for both Ovens using Logic Builder Logic in Routing to Oven_1 Logic at O1 Buffer Primer Oven_1 Load_Fixture O1 Buffer 31 31

3. Run Model for 5 replications 4. Export the Time Series data for both Variables from the Output Viewer to Excel 5. Name the File & Path then hit Export button 6. Copy the data from Excel into a Minitab worksheet Set File Name & Path 32 32

7. In Minitab, Stack the data into single columns 8. Select and run the Hypothesis Test using the data 33 33

Two-Sample T-Test and CI: Ov1_ALL, Ov2_ALL Two-sample T for Ov1_ALL vs Ov2_ALL N Mean StDev SE Mean Ov1_ALL 2250 95.5 22.0 0.46 Ov2_ALL 1870 66.2 17.9 0.41 Oven_1 Oven_2 Difference = mu (Ov1_ALL) - mu (Ov2_ALL) Estimate for difference: 29.312 95% CI for difference: (28.095, 30.530) T-Test of difference = 0 (vs not =): T-Value = 47.19 P-Value = 0.000 DF = 4116 Total Time at Ovens is NOT the Same Boxplot of Ov1_ALL, Ov2_All Individual Value Plot of Ov1_ALL, Ov2_All 140 140 120 120 100 100 Data 80 Data 80 60 60 40 40 Ov1_ALL Ov2_All Ov1_ALL Ov2_All 34 34

One-way ANOVA: Time versus OvN_RN Individual Value Plot of Time vs OvN_RN Source DF SS MS F P OvN_RN 9 878810 97646 238.43 0.000 Error 4110 1683215 410 Total 4119 2562024 S = 20.24 R-Sq = 34.30% R-Sq(adj) = 34.16% Time Total Time at Ovens is NOT the Same Individual 95% CIs For Mean Based on Pooled StDev Level N Mean StDev -----+---------+---------+- Ov1_R1 450 95.38 22.01 (*-) Ov1_R2 450 95.39 21.88 (*-) Ov1_R3 450 95.76 22.14 (-*-) Ov1_R4 450 95.54 21.94 (-*) Ov1_R5 450 95.62 22.04 (-*) Ov2_R1 370 67.86 18.57 (-*-) Ov2_R2 380 65.82 17.42 (-*-) Ov2_R3 370 65.80 17.91 (-*-) Ov2_R4 370 66.14 17.61 (-*-) Ov2_R5 380 65.53 17.88 (-*-) -----+---------+---------+- 70 80 90 100 Pooled StDev = 20.74 Time 140 120 100 80 60 40 140 120 100 80 60 40 Ov1_R1 Ov1_R2 Ov1_R3 Ov1_R4 Ov1_R5 Ov2_R1 Ov2_R2 OvN_RN Boxplot of Time Ov2_R3 Ov2_R4 Ov2_R5 Ov1_R1 Ov1_R2 Ov1_R3 Ov1_R4 Ov1_R5 Ov2_R1 Ov2_R2 OvN_RN Ov2_R3 Ov2_R4 Ov2_R5 35 35

The Big Y is Cycle Time per part Some Critical X s (things that could be changed to possibly reduce CT): Operation & Waiting Times Batch Sizes Capacities of Equipment Use of Resources Arrival of Parts Rework Rates Shop Layout 36 36

Define Potential Solutions Assess Potential Solutions Develop Proposed Solution Pilot the Solution Validate the Potential Improvement Define to be Process Map 38 38

Reduce Primer & Paint Booth Times Increase Capacities of Primer & Paint Booths Increase or optimize Oven Batch sizes Use extra, unused Oven 2 Capacity to help Oven 1 Increase Resources (X) Make quality changes to reduce rework % Reduce Inspections (X) Dedicate each Paint Booth to only 1 part color to reduce Setups Change Shop Layout to minimize or eliminate transporting parts Change Resource use for operation and / or transportation tasks Only Pull parts into the shop as new orders arrive (X) = Management says No! 39 39

1. Reduce Primer time by 50% 2. Reduce Paint time by 50% 3. Increase Primer Capacity by 1 4. Increase Paint Booths Capacity by 1 each 5. Increase Oven 1 Batch Size to 20 6. Reduce Oven 1 Batch Size to 10 (like Oven 2) 7. Increase Oven 2 Batch Size to 15 8. Change Oven 1 & 2 Capacities so Oven 2 helps Oven 1; (Oven 1 Cap. = 2 Batches, Oven 2 Cap. = 3 Batches) 9. Reduce Rework rate to 5% 10. Dedicate Paint Booths to 1 color each to eliminate setups 11. Pool Workers 2 & 3 for part moves 12. Pool All Workers for part moves 13. Use No Workers for part moves (in cellular shop) 14. Pool All Workers for All Tasks & Moves 15. Pool All Workers for All Tasks & No Moves (in cellular shop) 41 41

42 42

Note Average Time in System 43 43

You could build 5 additional Scenarios with the following Factor Combinations: S16 - Combination 1 = Scenarios: 15, 9, 3, 1, 4 & 2 S17 - Combination 2 = Scenarios: 14, 9, 3, 1, 4 & 2 S18 - Combination 3 = Scenarios: 13, 9, 3, 1, 4 & 2 S19 - Combination 4 = Scenarios: 12, 9, 3, 1, 4 & 2 S20 - Combination 5 = Scenarios: 11, 9, 3, 1, 4 & 2 44 44

Note Average Time in System 45 45

Combination 1 = Scenarios: 15, 9, 3, 1, 4 & 2 15: Pool All Workers Tasks with No Moves 9: Reduce Rework from 10% to 5% 3: Increase Primer Capacity from 1 to 2 1: Reduce Primer Time by 50% 4: Increase Paint Booth Capacity from 1 to 2 2: Reduce Paint Time by 50% 46 46

TP is ~520 Baseline WIP is slightly increasing 47 47

Baseline CT is better although still increasing slightly 48 48

Overall CT Averages ~249 Minutes 49 49

Entities are Blocked ~36% of the time vs. 68% in Baseline Entities are in Operation ~47% of the time vs. 19% in Baseline 50 50

Workers are Utilized ~70, 39 & 14% of the time 51 51

Scenario 16 Decreased Blockage & Waiting at Paint Buffers & Load Fixture Baseline 52 52

Paint Shop Parts are Stored after Key Process Steps 3. 1. Customer Orders Begin Here and Parts are Pulled as Needed Work Unit 2 Work Unit Work Order (Exit) (Exit) Entry Worker 1 Worker 3 WU Order WU2 Order Order Type WU Rw 50% Primer Oven_1 Load_Fixture Rework Inspection FG Buffer FG Storage Pull Orders WU2 Rw (Exit) Redo O2 Storage Worker 2 0000 O1 Buffer Number Completed Unload Fixture O1 Storage 0000 O2 Buffer Number Rejected 0000 Paint 1 Pull Order Time (minutes) Paint Storage Oven 2 0000 Work in Process Paint 2 Fixture Clean & Prep 0000 Cycle Time (minutes) Fixture 2. Replenishment Signals Order More Parts When Quantities Reach Trigger Level 53 53

Higher TP Than Original Baseline WIP Stabilizes 54 54

WIP goes up; however, it stabilizes! Cycle Time is higher; however, it also begins to stabilize! If Customer Work Orders arrive every 30 minutes then the amount of time to fill a Work Order drops to only 2 minutes (which is the actual pick time)!! Voice of the Customer / Voice of the Business Customer survey says they want CT to be <= 300 minutes Business is not sure if that CT can be met; however, they want orders quickly filled 55 55

Outgoing 10 8 Inspection Area Table Unload Fixture 7 6 Oven 2 Table 5 5 Paint 1 Table Paint 2 Door 9 Rework Area Bench Bench Supplies Worker 1 Worker 2 Worker 3 Incoming Roller Table Primer Table Oven 1 Table Load Fixture 1 2 4 Fixture Clean & Prep 3 Proposed Paint Shop Layout - 1 Door If Paint Booths are Monuments and/ or Extra Workspace is needed for other operations Table Oven 2 Bank of 3 Ovens Table Excess Property to be Removed 56 56

Implement Statistical Process Controls Determine Process Capability Verify Benefits, Cost Savings, Finalize Documentation 58 58

P rocess Data LSL 0 Target * USL 300 Sample M ean 247.121 Sample N 12575 StDev (Within) 59.7357 StDev (O v erall) 70.2831 Baseline had PPM > USL ~882K vcycletime (16 Comb1-15,9,3,1,4,2) - Capability Analysis (using 95.0% confidence) CT is More Capable But Needs Improvement O bserv ed Performance PPM < LSL 0.00 PPM > USL 107355.86 PPM Total 107355.86 LSL 0 150 USL 300 Exp. Within P erformance PPM < LSL 17.60 PPM > USL 188022.22 PPM Total 188039.82 450 600 750 900 Exp. O v erall Performance PPM < LSL 218.98 PPM > USL 225915.73 PPM Total 226134.71 1050 Within Overall P otential (Within) C apability Z.Bench 0.89 Low er C L 0.86 Z.LSL 4.14 Z.U SL 0.89 C pk 0.30 Low er C L 0.29 Upper C L 0.30 O v erall C apability Z.Bench 0.75 Low er C L 0.73 Z.LSL 3.52 Z.U SL 0.75 Ppk 0.25 Low er C L 0.24 Upper C L 0.26 C pm * Low er C L * 59 59

Simulation models can be a vital part of any LSS or other Process Improvement effort. DMAIC is a good approach for utilizing simulation models to help Define, Measure, Analyze, Improve, and Control any process. Process Simulator is a flexible, robust predictive analytics tool useful for process improvement! 60 60

Thanks for Attending! Any?? Instructor Info: Dave Tucker, LSSMBB ProModel Senior Consultant Office: 321.567.5642 dtucker@promodel.com Instructor Info: Bruce Gladwin, PMP, 6σBB VP, Commercial Products Office: 801.223.4639 bgladwin@promodel.com 61 61